public static void QSolve(string modelName = "n")
{
NGramModelBase model = NGramModelFileLoader.LoadByName(modelName);
var inputer = new NGramInputer(model);
inputer.PrintDistributeSize = 5;
Console.WriteLine($"Using model: {model.GetType().Name}");
while (true)
{
Console.Write("input > ");
var input = Console.ReadLine();
try
{
inputer.Clear();
foreach (string pinyin in input.Split())
{
inputer.Input(pinyin);
}
//Console.WriteLine("SubResults:");
//foreach (var result in inputer.SubResult.Take(5))
// Console.WriteLine(" " + result);
Console.WriteLine(inputer.Results.First());
}
catch (Exception e)
{
Console.WriteLine(e);
}
}
}
public static void QModel(string modelName)
{
NGramModelBase ng1 = NGramModelFileLoader.Load <NGram1Model>();
NGramModelBase ng2 = NGramModelFileLoader.Load <NGram2Model>();
NGramModelBase ng3 = NGramModelFileLoader.Load <NGram3Model>();
NGramModelBase ngn = NGramModelFileLoader.Load <NGramNModel>();
while (true)
{
Console.Write("data > ");
var input = Console.ReadLine(); // [char]+ [pinyin]
string chars = input.Split()[0];
string pinyin = input.Split().ElementAtOrDefault(1);
var condition = new Condition(chars, pinyin);
try
{
Console.WriteLine("1-gram");
ng1.GetDistribute(condition).Take(5).Print();
Console.WriteLine("2-gram");
ng2.GetDistribute(condition).Take(5).Print();
Console.WriteLine("3-gram");
ng3.GetDistribute(condition).Take(5).Print();
Console.WriteLine("n-gram");
ngn.GetDistribute(condition).Take(5).Print();
}
catch (Exception e)
{
Console.WriteLine(e);
}
}
}
public static void BuildModel(string filePath, IEnumerable <string> modelNames)
{
var stat = new TextStatistician(filePath);
foreach (var modelName in modelNames)
{
var model = NGramModelFileLoader.NewByName(modelName);
model.FromStatistician(stat);
model.Save();
}
}
public static void Solve(string inputFile, string outputFile, string modelName = "n")
{
var model = NGramModelFileLoader.LoadByName(modelName);
var inputer = new NGramInputer(model);
using (var outputWriter = File.CreateText(outputFile))
{
foreach (string input in File.ReadLines(inputFile))
{
inputer.Clear();
foreach (string pinyin in input.Split())
{
inputer.Input(pinyin);
}
outputWriter.WriteLine(inputer.Results.First());
}
}
}
public static void TestOnData(TestOption opt)
{
IEnumerable <NGramModelBase> models;
if (opt.StatFiles == null)
{
models = opt.ModelNames.Select(NGramModelFileLoader.LoadByName);
}
else
{
models = from stat in opt.StatFiles.Select(path => new TextStatistician(path))
from model in opt.ModelNames.Select(name =>
{
var model = NGramModelFileLoader.NewByName(name);
Console.WriteLine($"Built model: [{name}] from [{stat.SourceName}]");
model.FromStatistician(stat);
return(model);
})
select model;
}
var inputers = models.Select(model => new NGramInputer(model)).Cast <FullPinyinInputer>().ToArray();
var tester = new InputerTester(inputers);
using (var inputFile = File.OpenText(opt.InputFile))
{
if (opt.OutputFile == null)
{
tester.TestData(inputFile, Console.Out, opt.Format);
}
else
{
using (var outputFile = File.CreateText(opt.OutputFile))
tester.TestData(inputFile, outputFile, opt.Format);
}
}
}
public void TestNGramN()
{
TestOnData(NGramModelFileLoader.Load <NGramNModel>());
}
public void TestNGram123_Lambda()
{
TestOnData(NGramModelFileLoader.Load123_Lambda());
}
public void TestNGram12_MaxN()
{
TestOnData(NGramModelFileLoader.Load12_MaxN());
}
